Development of computational tools for cardiovascular applications

Cardiovascular diseases (CVDs) are a group of disorders of the heart and blood vessels and constitute the leading cause of death globally. Among the most common CVDs are pulmonary atresia and abdominal aneurysms which are mainly treated with modified Blalock–Taussig shunt operation (MBTS) and Endovascular Aneurysm Repair (EVAR) procedure respectively. During MBTS, a synthetic shunt is implanted between the subclavian and pulmonary arteries to supply the blood with sufficient levels of oxygen. Even though the shape of the implant plays a crucial role in the operation’s success, its’ selection relies on clinicians’ experience. To offer a personalized shunt morphology recommendation to patients, we developed a static Reduced Order Model (ROM) trained on Computational Fluid Dynamics (CFD) data which predicts the fluid flow solution as a function of critical morphological shunt parameters. The MBTS morphological parameterization was performed with the use of Radial Basis Functions. The ROM validation exhibited high precision, indicating a strong possibility of being integrated into clinical use. EVAR is a minimally invasive procedure for the treatment of abdominal aortic aneurysms, and it is usually performed in patients with aortic diameter exceeding 5 cm. Even though this operation is characterized by low mortality and morbidity rates compared to open surgery, it also comes with some complications such as challenging intraoperative navigation due to guidewire-induced deformations and significant risk of post-operative thrombus formation. Concerning those issues, we performed two distinct studies. The first work refers to the development of a pipeline based on ROM and Finite Elements (FE) simulations for the prediction of the aortic configuration in function of the guidewire orientation, the pre-operative aortic geometry, and material properties in almost real-time. Thanks to this tool, we were able to explore a wide spectrum of possible EVAR cases in a fast and efficient manner. The ROM prediction was found to offer sufficient precision. The developed framework showed the potential to support the clinicians’ decision-making process pre and intra operatively, reducing the use of radiation and contrast agents. The second work is related to the prediction of post-EVAR intra-prosthetic thrombus (IPT) formation through a holistic CFD analysis among simplified models of the three most popular commercial stent grafts, i.e., Zenith Alpha, Excluder, and Endurant II. For this work, we quantified the backflow, investigated the orientation of velocity vectors, and calculated the shear strain rates as well as the most common Wall-Shear-Stress related indices. Our main finding was that the simplified model of Zenith Alpha experiences higher recirculation and shear strain rates on its flared extension compared to the other grafts, which might explain its relatively higher susceptibility to thrombus formation. Conversely, the WSS parameters provided us with controversial results. In summary, three independent works were developed in the framework of this Ph.D. Two of them were focused on the ROM build-up to bridge the time gap between computational simulations and clinical reality. Achieving almost real-time ROM response on both works seems a promising start, which needs to be further investigated. Acknowledging the diversity in patients’ morphologies, both studies were conducted with a focus on individualized healthcare guidance through the consideration of patient-specific anatomies. The third study, dedicated to the postEVAR IPT formation, revealed that blood flow patterns, developed within a stent graft, can regulate thrombotic phenomena. The method of analyzing blood flow data was found to be crucial, showing that the conventional examination of WSS parameters might fall short of information. Finally, it appears that comprehensive and interdisciplinary approaches are essential and efficient both for explaining intricate cardiovascular issues and for planning and performing successful personalized treatments.